DE102015017039A1 - Automatic analysis of written and spoken texts of a source language (shown here using the example of the German language) in terms of their linguistic-grammatical consistency, which is an improved starting point for translations into other languages - Google Patents

Abstract

For more than half a century, so-called computer linguists have been trying to make natural language (including their grammar, logic, etc.) mathematically calculable. H. To formalize natural language. The goal is necessary, but it is tackled as if apparatuses and computers could solving all the associated problems on their own. The experience since 1957 (Chomsky), if not even since 1887 (Frege), teaches the opposite. The natural languages must first of all be formalized with their own possibilities to their limits. Only then are they mathematics a little bit more equal and may take advantage of mathematics a few advantages. The most obvious way was that of us newly developed: 1) The vocabulary of a language aufzugleisen enough that only one sense per word (including new artificial word) exists and 2) from the beginning, almost as rigorously and rigorously as it is in analogue form K In his second "Incompleteness Theorem" with mathematical statements and statements, Goedel did all his own numbers to number all terms and words and new and old meanings! On this basis, it is shown in this patent invention that even natural language grammars can be analyzed so comprehensively and comprehensively (ie, including all words) that they can be easily edited with numbers and automatically do things through numerical algorithms that the computational linguists do have not even begun to get started with their alien and fragmentary machines. Among the useful applications are: 1) An automatic grammatical analysis of self-language texts that can be used as the basis for translation engines, 2) Contextual recognition using a specific word field, in which, among other things, meaningful words are closely interlinked and 3) a direct one Translational possibility in foreign languages, where it is less important to the grace and beauty of the language, but rather to a literal, reliable and individual one-to-one translation, as is highly necessary in technical professions. 4) Other side effects are context recognition and thus correct word choice in texts, automatic summaries of large amounts of text and a new kind of speech recognition.

Description

This is an application known to the DPMA in another form but with exactly the same content since 28 August 2013 (file reference 10.2015.001.497.8, withdrawn on 22.01.2015).

Subsequently, the same invention was submitted again on 07.02.2015 and 17.02.2015 in two steps, once by fax and once by letter post, for which the Az. 10.2015.001.596.6 was issued.

This process to 10.2015.001.596.6 is still running. It will be referred to in this new application, which is intended to be a formal improvement, so that the old priority of 07/02/2015 can be applied for for this new application, which is submitted within the 12 month period.

The development of the word field with its context recognition (subclaim), as has been described in 10.2015.001.596.6, is not referred to again in this present application. However, the development of the translation engine and its preliminary stage, the textual analysis, will be discussed in more detail.

B)

Reason:

Even a human translator must always be aware of the meaning of the original text (source text) to be translated. Not only is knowledge of the vocabulary sufficient, which is rarely unique in natural languages. The interpreter always has to ask himself, what the author of the source text really wanted to say, because words are not unique, d. H. since they are usually occupied with several senses per word, are.

In order to ensure at least the first half of the translation process, namely the text recognition of the source text to be translated, most machine translations claim that they always "analyze" the source text (source text) first.

But this does not happen and can not happen without the idea developed here of assigning only one meaning to each word and only one number.

For how can clear, mutually delimited, exact and in any quantity available numbers be made compatible with the few existing and inconsistent and imprecisely defined natural language words? This only works if the language itself is mathematized, i. H. if she acquires some advantages of mathematics, d. H. if it splits at least so far that every word has only one meaning and a unique number for it, which can then also be used correctly.

C)

The patentability of the idea registered here:

• 1) Da sich diese hier vorgestellte Verbesserung auf Uebersetzungsmaschinen von Menschensprache bezieht, hat sie auf den ersten Bick mit nicht-patentfaehiger ”Sprache” und ”Methoden” zu tun. Tatsaechlich aber wird bei diesem hier vorgestellten Verfahren die gesamte Sprache konsequenter, als es je zuvor in der theoretischen Informatik geschah (also sehr viel weitreichender als in ASCII oder im Unicode oder in den programmiersprachlichen Feldern, bzw. Arrays und Strings oder in den Datenbanken usw.) nach vorheriger sprachlicher Analyse (also mittels einer umfassenden Sprachgrammatik, aber auf keinen Fall mittels einer angeblich formalen Grammatik mittels „Praedikatenlogik” und aehnlich Unbrauchbarem in nichts Anderes als in Zahlen ueberfuehrt, um tatsaechlich nur mit diesen Zahlen numerische Algorithmen erstellen zu koennen. Somit bleibt von der ”Sprache”, nach der anfaenglich noetigen, gruendlichen natursprachlichen Analyse fast nichts mehr uebrig. Denn die Loesung dieses oben genannten Problems ist rein technischer Art.
• 2) Auch als Konzept fuer neue ”Computerprogramme” sollte die hier vorgestellte Verbesserung fuer Spracherkennungsmaschinen und Uebersetzungsmaschinen keiner Patentierbarkeit im Weg stehen, da dieses Programm immer in technische Geraeten (zum Beispiel Sprachcomputern oder akustischen Steuerungsgeraeten) implementiert werden wird. Sie stellt also immer eine technische Loesung und Verbesserung eines technischen Problems von technischen Geraeten dar.
• 3) Darueber hinaus ist die hier vorgestellte technische Loesung, die der Industrie und dem Gewerbe zugefuehrt wird, neu.
• –) Neu ist die nach oben hin grenzenlose Aufsplittung des natursprachlichen Vokabulars.
• –) Neu ist, dass bei mehreren, voneinander abweichenden, Sinngehalten ein und desselben Wortes auch neue Woerter gebildet werden.

This patent application is about an improvement for all translation engines used so far, including the multi-patented (!) Translation machine "translator" operated by Google.

• 1) Since this improvement is related to translation of human speech, it has to do at first glance with non-patentable "language" and "methods". In fact, in this method, the whole language is more consistent than ever before in theoretical computer science (ie much more far-reaching than in ASCII or in Unicode or in the programming language fields, or arrays and strings or in the databases, etc.). ) after previous linguistic analysis (ie by means of a comprehensive grammar of speech, but in no case by means of an allegedly formal grammar using "Praedikatenlogik" and similar useless in nothing other than in numbers converted, in order to actually create numerical algorithms only with these numbers From the "language", according to the initially necessary, natural language analysis almost nothing left, because the solution of this problem above is of a purely technical nature.
• 2) Even as a concept for new "computer programs", the improvement presented here for speech recognition machines and translation machines should not stand in the way of patentability since this program will always be implemented in technical devices (for example voice computers or acoustic control devices). So it always represents a technical solution and improvement of a technical problem of technical devices.
• 3) In addition, the technical solution presented here, which is supplied to the industry and the trade, is new.
• -) What is new is the unlimited splitting up of the natural language vocabulary.
• -) What is new is that new words are formed in the case of several divergent, meaningful meanings of one and the same word.

• –) Neu ist also die Numerierung aller dieser oben genannten alten und neuen Woerter plus aller ihrer Nebenformen, Unterformen, Abweichungen, grammatisch bedingten Veraenderungen usw. Denn es werden nicht nur Grundwoerter numeriert, sondern alle Erscheinungsformen eines Wortes. Damit beinhaltet jedes einzelne Wort auch bereits seine grammatische Form.
• –) Diese Numerierung pro Wort erfolgt nach einem vorher einheitlich definierten Sinn, sodass es nicht mehr moeglich ist, einem Wort mehr als einen Sinn zuzuordnen.
• –) In dieser hier vorliegenden Entwicklung wird das Prinzip, nach dem ein Quelltext automatisch analysiert wird, anhand der deutschen Sprache dargestellt.
• –) Analog ist dasselbe aber auch mit jeder anderen Natursprache moeglich.

Although these do not have to be learned by any user of a natural language, they get their own unmistakable number in this system.

• -) New is the numbering of all of these old and new words above plus all their secondary forms, subforms, deviations, grammatical changes, etc. For not only basic words are numbered, but all manifestations of a word. Thus every single word already contains its grammatical form.
• -) This numbering per word follows a previously uniformly defined meaning, so that it is no longer possible to assign more than one meaning to a word.
• -) In this development, the principle of automatically analyzing a source code is shown using the German language.
• -) Analogously, the same is possible with any other natural language.

D)

The state of the art:

• 1) Chomsky's transfer method was doomed to fail. It went all the way down to a small art grammar, programming languages and a finite 5-tuple machine offensive to any algorithm that can detect its changing states. After that, the development was over. Such primitive programs, as prepared by Frege, Whitehead, and Russell almost 150 years ago with their predicate logic, are not suitable for natural language translations or even just analyzes of source texts.
• 2) Even the so-called Interlingua method, which tried to translate into intermediate texts, has never been able to produce useful results.
• 3) The two newer methods EBMT, which (apparently) "translate" on the basis of often existing sentences and frequently occurring word arrangements, and the SBMT mentioned at the outset, which requires large sample quantities for their (alleged) "translation" as templates, are intelligent " Translation capabilities ", which in this way (in the EBMT statistics and in the SBMT sentence examples from the in-house database) avoids a real analysis of the source texts to be translated.

However, it is the described reasons (namely that the approximate exchange of similar text modules and common phrases is not a reliable and meaningful "translation"), why even in these last two cases usable and reliable translations can not succeed, as long as not before a comprehensive treatment of Languages in the way we have developed and set out.

e)

• 1) Bildlich gesehen, kaempfen translation machines, die die Uebersetzungsprobleme mittels Einspeisung grosser Textmengen zu loesen versuchen, ”an zwei Fronten”, naemlich an der Front einer passenden und verstaendlichen Uebersetzung im Zieltext plus zusaetzlich an der Front einer richtigen Erkennung des Quelltextes.

Why this idea presented here is needed for all (natural language or foreign language) natural language (output) or controlled (input) machines:

• 1) Figuratively seen, translation machines that attempt to solve the translation problems by injecting large amounts of text are fighting "on two fronts", namely at the front of a suitable and comprehensible translation in the target text plus the front of a correct recognition of the source text.

Assuming that such a TM would have solved the first-mentioned problem of a good speech representation in the target text by means of its many fed (flawless) sample texts, it can nevertheless Do not reliably come to a correct correspondence in the respective source text, if not previously an analgesic in the manner described here has been made.

In other words, in such a case (a non-interpretation of the source text beforehand) even a good TM with its translations would always move on an unsecured ground, ie have two uncertainty factors at the same time. This problem will persist (with the exception of blanket exchangeable phrases, standard phrases, etc. which, as defined here, have nothing to do with sophisticated translations), as long as there are too few words per colloquial language and as long as people with these few, imprecise terms (all in addition their own languages) can arbitrarily juggle around and rearrange them in endless combinations over and over again.

Translation machines such as those described above, like human translators, would also have served with the "translating aid" developed here (or heretofore) if they were to overrun their source texts as a sort of corset, only then, on this then consolidated basis to do their more or less free translations.

For on the page of the target text, variable translations as offered by Google do not necessarily violate the intended meaning of the source text, as long as they build on the secure ground of a precisely analyzed source code. But if both sides (source code capture and target text reformulation) are indefinite and "mobile", it will be difficult to establish reliable and, above all, self-controlling links between the two.

F)

Implementation:

1) Theoretically:

The approach presented here begins with the new difference that the own database is not constantly fed with finished sentences or even whole texts (as with the mentioned "stastistic based" translation engines), but is built up only with individual words. Also, the point here is not, by probability, to filter out the most-used words, and in wordings, the most closely worded word combinations, but to separate each word from just similar words, and for each conceivable subform of that word, its own word (with its own place and name) own number) in its own dictionary of words.

The different meanings of a word are roughly split up by (for example, German) dictionaries, in which homonyms are listed separately and made aware of the slight differences in meaning by means of phrases. So, obviously, there will "give itself the word", "keep word", "a man, a word", "in the beginning was the word" and similar distinctions. A few more examples provide foreign language encyclopedias.

But the here to be developed and regularly further to build up books of books in addition to these o. G. different meanings and some other synonymous still all grammatical forms, so that each grammar form is a separate word with its own number and thus each word is defined grammatically unmistakable.

Apart from a few "irregular" grammatical forms, most of the grammatical subordinate forms of the basic words (which in total can extend the volume of the vocabulary up to 500 times) are not produced manually but algorithmically (see Appendix 1).

If such words with their grammatical additional information together as the correct, d. H. If the words belonging to the words appearing in the source code are identified, the first requirement for a comprehensive grammatical "analysis" has already been fulfilled. In the ideal case, the grammatical coherence thus found even allows the recognition and reporting of incorrect grammatical applications in the source text.

It will also be advantageous for acoustic speech recognition if structured sentences with precisely defined, individually recorded, words and syllables can be more accurately identified in a recognition computer than whole word collections or sentences in which the words are or word syllables can become interlinked and confused by inexact or fast pronunciations.

2) Practical:

The sentence "this is the house" contains 4 words. Through careful analysis of their meaning, even these four simple words produce new or slightly deviant new words, which receive their own integers.

The word "das" is a word that despite identical appearance is occupied by different words. To name just two of them: The first "that" is a pronoun, the second "that" is an article. These two words are different words in meaning and thus get their own integers independently of each other. This diversity also shows the English parallel sentence "that" (not "the") is the house ".

The same applies to the word "is". Also for this word, there are many different meanings that require their own number assignment. To take only the two most important examples: The one "is" could be a Hilfszeitverb, the other a Hauptverb.

If a grammatical and sense analysis of a source text were to be checked for correctness or to prepare it for translation into another language, the words from the source text would be word for word attached to one's own, constantly growing, vocabulary and often have several same-looking words with different numbering.

In this simplified example above, they each traverse two different words. At the first "that" the program would not know, if this is an article or a pronoun. Following a noun, it is clear that the previous "that" was an article and matches the case and gender of the noun "house".

Another example is attached in Appendix 2.

This is the first way out of natural language inaccuracy due to missing words and inconsistent or missing word definitions and confusing homonyms, etc., and from the differences between the different languages, all of which have the same problems but in different parts of the vocabulary, so that recognized vulnerabilities of one Word in one language can not be translated to that word in another language.

Why does a real number-formalization require prior analysis with grammatical means?

There is not a single word in (German) language that is grammatically not defined in an unmistakable and unique way. This is why the analysis of the source text employs speech grammar, not the reduced and incomplete formal grammar. Correspondingly, all possible words and word forms can be provided with their own, unique numerical code, which represents a unique grammatical description (and can later be assigned one-to-one in translations).

Although the construction of such a lexicon up to a 500-fold extension of the underlying basic word stem, the additional emerging new words and the grammatical subforms must not be entered manually into the encyclopedia to be developed here, but can be formed using algorithms. Also exceptions, such. As the irregular verbs, would be calculable by algorithms and entered into your own lexicon. However, they represent such small quantities that they can also be entered manually into the lexicon developed here.

Its own dictionary of words is a relational database that is easy to manage because only integers (ie, about 2.1 billion integers) appear as a data type. Only after the first key attribute, ie the one number, which is assigned to a single word in the lexicon, does this word appear as a character and in alphabetical order, so that the written words of the source text can recognize their counterparts.

The remainder, ie all contents of the columns consist only of numbers, namely of arbitrary additional codes (which specify something like "transitive verb", "abstract noun" and others) and from the numbering of other words that are used as foreign keys with the existing data set, because they typically appear in source texts and serve the present word in its grammatical and other analysis.

In the one-to-one splitting of a vocabulary (ie there is only one meaning per word), there are always very few double assignments, of which the additional and environment words listed in the data record are easily the correct word that is identical to the source text , can be identified.

Even the many hundred thousand verbs with their more than 500 forms of conjugation can be "formalized" at a normal natural language level by means of "quite normal" natural language grammar (which obviously does not exist for the mathematical logician, because it has not even been attempted to use it until now). d. H. to numerical algorithms. Thus verbs, including passive, despite their 500 conjugation forms in the diffraction cause only 16 different endings and are therefore easy to manage and identify in their predictable order (see Appendix 2).

G)

As a side effect of this new approach presented here is also a completely new and first meaning-accurate translation engine (the project is internally called "HANA I"):
If two (or more) vocabularies of the same numerics (including the same one-to-one meaning allocation) are used, so that two (or more) languages are compatible with each other via the common numerics, then the equivalent words are those multilingual vocabularies interchangeable one-to-one. This requires only for the natural language source code of the chars as a data type in the first column of each record, but otherwise all columns are filled only with numbers, ie mainly with the "foreign keys" other word numbers, as environment words in the source code are mostly typical and belong to the word you are looking for, whereby algorithmically only the same numbers have to be exchanged in order to exchange the same words or their literal meaning.

It would remain with such an exchange in the word order of the source text in the target text, the unchanged word position from the source, this could in the further use of the delivered text, either a) is a ready translation of the translation engine developed here or b) from the back translation of Source text in their own language was created, but cause no grammatical or meaning-related damage in the target text. Because this damage would be compensated by the text modules of the finished word arrangements of the respective target languages, as they use other machine translations, as all tests prove.

The advantages of translations by the newly presented translation engine (HANA I) are indeed in their one-to-one translation, ie in their "literal" accuracy and are therefore particularly suitable for technical and scientific texts in which it applies to each individual Word, but if these translations of the machine presented here are to be made more elegant and easier to read by means of other translating machines which follow the described principle of exchanging entire sentences or phrases, these translation machines can subsequently make the above translations of " HANA I "once again as the source text for a second translation into the same target language.

If "HANA I" only provides a translation into own language (eg into German), this translation must be done using another translation engine. If a foreign language translation is delivered by "HANA I", this further translation can be made optionally.

Effective is such (be it eigentextliche or foreign language) translation, which provides exactly the right terms and their precise grammatical analysis by means of the translation engine "HANA I", for the re-users, but only if the recipient language is used per sense of a Words exactly the same numbering served as the delivering here developed translation engine, because only the numbering of the individual words, the analysis of the delivered text can be seen. There would be little recognizable in the natural language words themselves.

As already mentioned, for example, the word "house" without an article is not to be seen, which case it dresses. In the numbering, however, the "house" in the 1st case has a different number than the house in the 4th case.

H)

General reason why language programs in computers are part of a technical process:
Not only the machine language of the computer, but also the human natural language can (since computer participate in technical processes!) Be part of the technology.

Since humans also become part of technology within technical processes (or by definition to "natural forces" or "technical means"), and since their human natural language is generally more widely available, this human language is also one of the above. technical processes and technical solutions.

The machines (including computers) communicate with other machines (including computers) via their own art and machine languages, but where people appear in the communication chain, they also need to be communicated through the diverse human natural languages.

• 1) Der Mensch spricht und die Maschine agiert oder schreibt: Der bekannteste Fall ist der von St. Hawking, dessen Computer das schreibt, was er akkustisch eingibt. Dabei arbeitet dieser sog. Sprachcomputer sogar mit einem „statistic based”-Programm, das aus seinen bisherigen Formulierungen weitgehend individuelle Formulierungen generiert.

The latter concerns both spoken and written communication in both directions: speaking and listening or writing and reading.

• 1) Man speaks and the machine acts or writes: The best known case is that of St. Hawking, whose computer writes what he inputs acoustically. In the process, this so-called speech computer even works with a "statistic based" program, which generates largely individual formulations from its previous formulations.

Other phonetic applications are in the field of safety engineering (eg identification), the control of machines and (industrial) robots and the like. v. a.

Are also known on computer driven speech recognition programs in z. B. Healthcare. In doing so, doctors do the fastest, because verbal, kind do their daily paperwork.

At the same time, initiatives are underway (eg the ADOK project as a bridge between engineering and German as a foreign language) or initiatives by the physicians' chambers are required to liberate foreign doctors in German hospitals from their language incompetence.

In the meantime there are various manufacturers (eg of the company ECTACO) of portable "compilation computers" of all sizes who transfer written and even spoken (ie acoustically recorded and spoken) language into the desired foreign languages. However, the experts mentioned are not served with such "text modules" and "phrases".

• 2) Der Mensch hoert und die Maschine spricht: Hier bekommt der Mensch die fuer ihn notwendigen Informationen (z. B. Navigation, Gefahrenhinweise, Instruktionen, Qualifikationsmassnahmen, Oberflaechenbedienung usw. betreffend) akustisch uebermittelt. So hat der Mensch seine Haende und den Blick frei, sowohl, wenn er der Letzte in der Informationskette ist, als auch, wenn er, sehr viel haeufiger, ein Verbindungsglied eines technischen Prozesses ist.

But there are also professional translation computers, which (analogous to the implementations in the ABS brake system) welcome an implementation of our developments.

• 2) The human being hears and the machine speaks: Here, the person receives the necessary information (eg navigation, hazard warnings, instructions, qualification measures, surface operation, etc.) transmitted acoustically. Thus, man has his hands and eyes free, both when he is the last in the information chain, and when he is, much more often, a link in a technical process.

Attachment 1

Concrete example of how a natural language text is analyzed with a natural language grammar and prepared for algorithmic processing:
Also, the following, by nature not complicated, example shows how with each text speech recognition and language translation can be made with natural language means and that the metagrammatics of formal linguistics with their "atomic" desired laws should include, inter alia, once the past.

• (1) Christoph Nomen (Eigenname), Eigenname, wahrscheinlich im Genus Masculinum. Falls nicht als solcher gespeichert, wird das Originalwort aus dem Quelltext in die Uebersetzung uebernommen. Als Eigenname von einem (meist menschlichen) Lebewesen ist dieses Wort fast immer im Numerus Singular und stellt ein artikelloses Substantiv dar, meist ohne Artikel, wenn dieses Nomen (meist im Plural) nicht gerade ein Appelativ (also eine Gattungsbezeichnung) ist. Der Kasus ist hier (noch) unbekannt. Spaeter wird dieses Nomen Dativ.
• (2) Kolumbus Nomen (Eigenname), (Eigenname), Dativ. Falls nicht als solcher gespeichert, wird das Original aus dem Quelltext in die Uebersetzung uebernommen. Zwei Eigennamen hintereinander machen es (im Deutschen) wahrscheinlich, dass es sich um Vornamen und Familiennamen handelt. Ausserdem um ein Singular. Alles Weitere ist hier wie unter 1). Falls jedes Wort des vorliegenen Textes (wie vorgesehen) mit jedem anderen Wort dieses Textes ueber das eigene Wortfeld verbunden ist, und erst recht, wenn schon bei der Speicherung dieses Namens zuvor nahe Begriffe in das Wortfeld eingegeben sind (also auch „Ei”, „Amerika” usw.), wird die Sinnerkennung zusaetzlich gesichert. Der Kasus der beiden Namen ist, wie gesagt, an dieser Stelle noch unbekannt: Da bei Eigennamen meist kein Artikel verwendet wird, sind hier fuenf Faelle moeglich, Vokativ („Christoph Kolumbus, komm mal her”) plus 4 weitere Faelle Singular: Der, des, dem, den. (der CK tut etwas, des CK Haus steht hier, dem CK geben wir etwas, den CK fragen wir etwas). Spaeter im Text stellt sich dieses Nomen als Dativ heraus.
• (3) wird Verb (Hilfsverb), 3. Person, Singular, Presens (vgl. „vorgehalten”). Nach dem Eigennamen Christoph Kolumbus ist zwar noch der Kasus dieses Substantivs unbekannt, aber es wird nach diesem Nomen ein Verb erwartet, das hier in Form eines „wird” auftaucht. Allerdings: a) Dieses „wird” kann bedeuten, dass CK (Nominativ) z. B. jetzt Vizekoenig „wird”. b) Dieses Wort „wird” kann (ebenfalls im Nominativ) aber auch ein Hilfszeitverb darstellen, um Futur I und II einzuleiten: CK „wird” Amerika entdecken (oder wird es entdeckt haben). c) Dieses „wird” kann aber auch zu einem Passiv gehoeren, wobei CK ebenfalls im Nominativ steht (CK wird belohnt oder wird belohnt werden). d) Es kann aber auch ein Passivhilfsverb sein (in 3. Pers. Sing.), das in Verbindung mit einem noch ausstehenden Partizip II eines Vollverbs eine sog. Partizipialphrase bildet und diesen ersten Hauptsatz (als 1. Wuerfel) abschliesst, bei der CK im Praesenz oder im Futur ein Dativ darstellt (z. B. dem CK wird etwas geschenkt oder ihm wird etwas geschenkt werden).
• (4) nach Praeposition (temporal, mit Dativ) Nach dem o. g. „wird” folgt als 4. Wort das Wort „nach” (es ist das erste Wort im zweiten Wuerfel). Entsprechend dem hier vorgesehenen Prozedere bleiben die oben genannten 3 Woerter („Christoph Kolumbus wird...”) in Wartestellung, und die naechsten drei nach dem Sinn erfassten kleineren Textteile bzw. „Wuerfel” (Einschuebe wie „nach seiner Rueckkehr aus Amerika” und „waehrend eines Essens bei Kardinal Mendoza” und „im Jahre 1493”) werden Wort fuer Wort nacheinander uebersetzt und wuerfelweise abgeschlossen. Diese o. g. genannten Wuerfel 2, 3 und 4 sind keine Aktanten (also semantisch bedeutsame Teilsaetze bzw. Satzglieder), sondern nur zeitlich und grammatisch nebeneinander gleichwertige Einschuebe innerhalb des erstgenannten, ihnen uebergeordneten, „Matrixsatzes”, und man kann schnell feststellen, dass keines ihrer Woerter dem erstgenannten Hauptsatz („Christoph Kolumbus wird”) irgendetwas bieten kann, um ihn zu ergaenzen und damit grammatisch und semantisch abzuschliessen. Praepositionn wie „nach” oder „aus” u. a. bilden keinen Satzteil, sondern fordern eine Ergaenzung. Es gibt aber viele Woerter „nach”, die im Deutschen homonym sind (was nicht heisst, dass sie es in der Fremdsprache auch sein muessen: Vgl. „sailing to Amerika” fuer „Segeln nach Amerika” oder „after lunch” fuer „nach dem Essen” usw.). Die Ergaenzung, nach der die Praeposition verlangt, entscheidet also erst ueber die Art des Wortes „nach” (die statisch lokal (nach der Ampel), zielgerichtet lokal (nach Amerika segeln), temporal (nach dem Essen), modal (nach Schweizer Art) u. v. a. sein kann). Erst das folgende Substantiv (oft erst hinter seinem Artikel), auf das sich die Praeposition bezieht, gibt der TM die Moeglichkeit, das richtige „nach” aus seinem Morphem-Vokabular auszuwaehlen. Hier im vorliegenden Fall folgt der Praeposition „nach” der Possessiv-Artikel „seiner” und dann das Sustantiv „Rueckkehr”.
• (5) seiner Possessivartikel, Femininum, Dativ, Singular („vgl. „Rueckkehr”). Was den Possessivartikel „seiner” betrifft, so schwebt bezueglich der Genera ueber allen Textteilen das grammatisch festgelegte Maskulinum des Cristoph Kolumbus (darum ist hier nicht von „ihrer” die Rede), aber die Kasus werden natuerlich vom Dativ Singular der „Rueckkehr” bestimmt, darum ist im vorliegenden Text von „nach seiner Rueckkehr” die Rede.
• (6) Rueckkehr Nomen, Femininum, Dativ, Singular Diese hier voliegende Interpretation ist einfach, denn das Nomen „Rueckkehr” ist eindeutig (d. h. es gibt keine Mehrdeutigkeiten bzw. Doppelbedeutungen dieses Wortes). „Rueckkehr” ist hier temporal, d. h. es fodert das zeitliche „nach” (und z. B. nicht das lokale). Auch dieses zeitliche „nach” fordert von seiner Ergaenzung ein Dativ. Dieser Dativ ist dem Hauptwort „Rueckkehr” nicht anzusehen, da „Rueckkehr” nur im Singular existiert und es im Dativ Singular kein Flexionsmorphem aufweist. Alle Kasus dieses Wortes (das eine Konversion des Verbs „zu-rueckkehren” darstellt) heissen immer nur „Rueckkehr”, ohne ein einziges Suffix am Ende. Pro Morphem kann das feminine Nomen ”Rueckkehr” 4-mal vorkommen (d. h. in allen 4 Kasus im Singular).
• (7) aus Praeposition (lokal mit Dativ) Bei „aus Amerika” handelt es sich ebenfalls um eine Praeposition („aus”) und ihrer Ergaenzung („Amerika”), einem lokalen Eigennamen, entsprechend meist ohne Artikel und immer im Singular. Was die anfaenglichen drei Woerter betrifft („nach seiner Rueckkehr”), so duerften diese abgeschlossen sein, denn das Wort „aus” leitet eindeutig eine neue Aussage ein, sodass nicht damit gerechnet werden muss, dass sich die drei scheinbar zusammengehoerenden Woerter „nach seiner Rueckkehr” noch einmal in ihrem Sinn veraendern muessen, weil nachfolgende Begriffe eine andere Interpretation dieser drei Morpheme fordern. Auch „aus” hat (genau wie „nach”) viele Bedeutungen („aus Versehen”, „aus der Mode”, „aus dem Gedaechtnis”, „fertig aus”, „ich frage ihn aus”, „aus Holz”, „aus Amerika” usw.). Es muss also fuer die Wahl des richtigen Wortes „aus” durch die TM vorher geklaert werden, um was fuer ein Wort „aus” es sich handelt.
• (8) Amerika Nomen (Eigenname), Neutrum, Dativ, Singular. Wieder gibt erst das folgende Substantiv Aufschluss. (Eigenname). „Amerika” benoetigt ein lokales „aus”, das einen Dativ verlangt, d. h. Amerika steht im Dativ. Dieser Dativ ist schwer zu erkennen, weil Amerika im Dativ Singular (genau wie bei der „Rueckkehr”) kein Suffix aufweist und (im Gegensatz zur „Rueckkehr”) keinen Artikel hat. Aber letzteres ist irrelevant, da die hier gegebenen zwei Woerter („aus Amerika”) eine aeusserst einfache Anordnung von Woertern darstellen und der ebenfalls einfach strukturierten Anordnung der ersten drei Woerter („nach seiner Rueckkehr”) nicht „in die Quere kommen”. Damit ist der zweite Wuerfel „nach seiner Rueckkehr aus Amerika” ebenfalls abgeschlossen.
• (9) waehrend Praeposition (lokal mit Genitiv) Das Wort „waehrend” ist nicht grundsaetzlich eine Praeposition. Hiesse der Satz im Text „waehrend sie assen”, waere „waehrend” nur eine sog. Subjunktion, die einen abhaengigen Teilsatz (Nebensatz) mit dem Hauptsatz verbindet. Zur Praeposition gehoert (wie bei den Praepositionen „nach” und „aus” bereits dargelegt) ein Substantiv, dessen Kasus (wie ebenfalls schon ausgefuehrt wurde) immer von der jeweiligen Praeposition bestimmt wird. Hier ist es der Genitiv (weniger korrekt wird hier aber auch manchmal der Dativ verwendet).
• (10) eines Artikel (indefinit), Neutrum, Genitiv, Singular („vgl. Essens”). Indefiniter Artikel, grammatisch gleiche Form wie „Essens” (Genitiv, Singular, Neutrum).
• (11) Essens Nomen, Neutrum, Genitiv, Singular, mit Suffix. Zwischen dem (aus dem Verb „essen” gebildeten) Substantiv „Essens” (Genitiv, Singular, Neutrum) und der Praeposition „waehrend” steht der infinite Artikel des Substantivs „eines” (ebenfalls im Genitiv, Singular, Neutrum). Pro Morphem kann das neutrale Nomen ”Essen” 3-mal vorkommen (d. h. in 3 Kasus im Singular).
• (12) bei Praeposition, (lokal mit Dativ) Auch „bei” ist wieder eine lokale Praeposition, die vom dazugehoerigen Substantiv den Normalkasus Dativ fordert und hier einen Raum bzw. eine Lage angibt.
• (13) Kardinal Nomen, Maskulinum, Dativ, Singular (zu „Mandoza”) „Kardinal” ist ein Titel, der normalerweise einen Artikel verlangt („bei dem Kardinal” oder „beim Kardinal”), hier aber als Bestandteil des Eigennamens „Mendoza” auf einen Artikel verzichten kann.
• (14) Mandoza Nomen (Eigenname), Maskulinum, Dativ, Singular (zu „Kardinal”) Da weder der Eigenname noch der Titel ihren Dativ per Artikel oder Wortendung zeigen, ist dieser Dativ unsichtbar und kann nicht an die Praeposition „bei” zurueckgemeldet werden. (Eigenname), Maskulinum, Dativ, Singular. Dennoch wird die TM diese Ungewissheit akzeptieren, da nichts dagegen spricht, dass es sich bei „Kardinal Mendoza” um einen Dativ (Singular, Maskulinum) handelt.
• (15) im Praeposition (hier temporal mit Dativ) Auch die folgenden 3 Woerter „im Jahre 1493” sind lediglich ein weiterer Einschub und nehmen grammatisch keinen Bezug auf die ersten 3 Woerter. Diese 3 folgenden Woerter beginnen mit der Praeposition „im” (wobei ein „in” mit dem folgenden definiten Artikel „dem” verschmolzen wird, was bei Abstrakta und vor allem Redewendungen (den sog. festen Wendungen) geschehen kann), d. h. einer Praeposition, die bei statischen Aussagen vom neutralen „Jahr” den Dativ („in dem (im) Jahr”) und bei gerichteten Aussagen den Akkusativ (bis in das (ins) Jahr”) fordern.
• (16) Jahre Nomen, Neutrum, Dativ, Singular, mit Suffix Statt „in dem Jahr 1493” hat sich also nicht nur die Verschmelzung zu „im” durchgesetzt, sondern auch die Formulierung „im Jahre”. Tatsaechlich ist das Suffix „e” hinter „Jahr” so unnoetig wie die ganze Formulierung „im Jahr” oder „im Jahre”, es wuerde auch genuegen, nur die Jahreszahl anzugeben. Aber, wie gesagt, es handelt sich hier um eine sog. „feste Wendung”, die als solche auch zusaetzlich zu den einzelnen Morphemen „im” und „Jahre” im Woerterbuch abgespeichert werden kann. Nach „im Jahre” folgt eine Vakanz fuer eine Erklaerung oder eine Jahreszahl.
• (17) 1493 Zahl (Jahreszahl) Die Zahl „1493” muss weder fuer die Sinnerkennung, noch fuer Uebersetzungen definiert werden.
• (18) vorgehalten, Verb, (Partizip II), (vgl. „wird”). Nach 13 Morphemen und einer Zahl, die alle einzeln und zusammen nicht geeignet gewesen waren, den anfaenglichen Hauptsatz („Christoph Kolumbus wird...”) abzuschliessen, erscheint nun aber an Position 18 das Wort „vorgehalten, Mit diesem Hauptverb „vorgehalten” wird der anfaengliche Hauptsatz endlich abgeschlossen und entscheidet nun erst ueber den Kasus von CK ganz am Anfang. Christoph Kolumbus steht dort also, wie man jetzt nach dem 15. Wort weiss, im Dativ, denn ihm wird „vorgehalten”. Vorhalten hat zwei Bedeutungen, d. h. dem CK wird eine Waffe, also ein konkreter Gegenstand, vorgehalten oder dem CK wird (wie es hier der Fall ist) etwas Abstraktes vorgehalten. Aber diese Unterscheidung ist grammatisch irrelevant. Wenn dem CK eine Waffe oder ein Verrat vorgehalten wird, kann in beiden Faellen der Satz mit einem dieser beiden Akkusativa beendet sein. Aber das Nichtvorhandensein eines Akkusativs schafft eine sog. Vakanz, d. h. es steht eine Erklaerung aus, was dem CK denn vorgehalten wird. Dieses ergibt sich daraus, dass der Aktant „(dem) CK wird vorgehalten”) zwar ein eigener Wuerfel sein kann, aber grammatisch noch unfertig ist. Beides koennte, wie oben erwaehnt, kurz mit einem Akkusativ geschehen: Dem CK wird eine Waffe bzw. ein Verrat vorgehalten (Punkt, Hauptsatz beendet). Oder dieser erwartete Akkusativ bleibt aus, dann wird stattdessen eine Erklaerung erwartet (die erklaert, was dem CK vorgehalten wird). Diese wird mit „dass” eingeleitet oder erfolgt in direkter oder indirekter Rede. Die Erklaerung folgt in diesem Text mit den Worten „es sei ein Leichtes gewesen” (also im Konjunktiv der indirekten Rede), die sodann eine weitere Vakanz erzeugt, indem sie auf einen untergeordneten Nebensatz wartet, der wiederum erklaert, was denn genau „ein Leichtes” gewesen sei usw. Diese hier voliegende Interpretation ist einfach, denn das Nomen „Rueckkehr” ist eindeutig (d. h. es gibt keine Mehrdeutigkeiten bzw. Doppelbedeutungen dieses Wortes). „Rueckkehr” ist hier temporal, d. h. es fodert das zeitliche „nach”. Auch das zeitliche „nach” fordert von seiner Ergaenzung ein Dativ. Dieser Dativ ist dem Hauptwort „Rueckkehr” nicht anzusehen, da „Rueckkehr” nur im Singular existiert und es im Dativ Singular kein Flexionsmorphem aufweist. Alle Kasus dieses Wortes (das eine Konversion des Verbs „zu-rueckkehren” darstellt) heissen immer nur „Rueckkehr”, ohne ein einziges Suffix am Ende.
• (19) es Pronomen Nach dem Verb des Hauptsatzes „vorgehalten” (Praesens Indikativ Passiv) und vor diesem „es” (mit dem ein neuer Satzteil beginnt, steht ein Kommma, das (da im Hauptsatz kein Nomen Objekt fuer die ausstehende Erklaerung vorhanden war) mit „dass”, mit direkter Rede oder (wie hier) mit indirekter Rede, die Erklaerung liefert, was dem CK vorgehalten wird. Hier handelt es sich um ein anderes „es” als z. B. bei „es ist mir schlecht”, einem Satz, bei dem das „es” ohne weiteres weggelassen werden koennte („mir ist schlecht”), da dieses vorgenannte „es” ein sog. Platzhalter ist und eine Funktion als Subjekt erfuellt. Bei „es sei ein Leichtes gewesen” (oder vereinfacht: Es sei leicht gewesen) darf das „es” nicht fehlen, weil genau dieser Akkusativ (es, dies, das usw., der im Folgenden ja noch naeher erklaert wird, naemlich mit „die „Neue Welt” zu entdecken”) als Subjekt zum Satz gehoert (einem Nebensatz 1. Grades, der grammatisch im Konjunktiv II der indirekten Rede unmittelbar den Hauptsatz (C.K. wird vorgehalten”) nachgeordnet ist.
• (20) sei Verb (Hilfsverb), 3 Person, Konjunktiv Der Nebensatz 1. Grades „es sei ein Leichtes gewesen” verwendet hier die indirekte Rede mit Referatskonjunktiv II. Ohne grammatischen Fehler haette das Hilfszeitwort „gewesen” auch wegfallen koennen, die Aussage haette dann, ebenfalls grammatisch korrekt, gelautet, „es sei ein Leichtes, die „Neue Welt” zu entdecken”. Semantisch genauer ist es aber, das Wort „gewesen” hinzuzufuegen, da es im Praesens und Futur keine Moeglichkeit der Wiederholung gibt. Denn die „Neue Welt” konnte nur einmal entdeckt werden.
• (21) ein Artikel (indefinit), Neutrum, Nominativ, Singular, (vgl. „Leichtes”). Hier handelt es sich nicht um eine Zahl, ein Kardinalzahladjektiv o. a., sondern um einen unbestimmten, infiniten Artikel. Auch ein unbestimmter Artikel wird mit dem dazugehoerigen Nomen mitdekliniert. Da das Nomen an dieser Stelle noch nicht bekannt ist, koennte das Wort „ein” entweder als Maskulinum Nominativ Singular vorkommen (ein Mann) oder als Neutrum Nominativ Singular (ein Haus) oder als Neutrum Singular Akkusativ (ein Haus).
• (22) Leichtes Nomen, Neutrum, Nominativ, Singular (vgl. „ein”). Es handelt sich hier um ein substantiviertes Adjektiv im Nominativ Singular. Fuer die hier zu entwickelnde TM ist es folgerichtig, dieses „ein Leichtes” woertlich zu uebersetzen, wenngleich mit fast dem gleichen Sinn auch nur das Adjektiv „einfach” ausgereicht haette.
• (23) gewesen, Verb (Partizip II), Infinitiv, mit Wortstammaenderung und Praefix. Wie unter „sei” dargelegt, handelt es sich bei diesem Konjunktiv um einen Konjunktiv II, also um ein Praeteritum von „sein”, bei dem dieses „sein” als finites Verb verwendet wird („sein” als Hauptwort) und seinen Perfekt, Plusquamperfekt und Futur II mit „gewesen” bildet (es ist gewesen, es war gewesen, es wird gewesen sein).
• (24) die Artikel (finit), Femininum, Akkusativ, Singular. Das vorige Wort „gewesen” und dieses naechste Wort „die” werden korrekterweise von einem Komma getrennt. Die in der TM entwickelten Definitionen, die hier in dieser Anlage nur z. T. an einigen grammatischen Regeln dargestellt wurden, sind derart zwingend, dass sie, auch bei falscher oder voellig fehlender Interpunktion funktionieren! Der einzige Nachteil waere im Falle falscher oder fehlender Interpunktion lediglich, dass die Sinnkontrolle ihre Arbeit erst ganz am Ende des Satzes beendet haben wuerde, wenn alle Woerter des Satzes trotz der fehlenden Interpunktion als richtig zugeordnet (oder auch nicht) erkannt worden sind. Das gesamte jetzt folgende Satzgefuege „die „Neue Welt” zu entdecken”, ist ein Nebensatz 2. Grades (der von dem vorherigen Nebensatz 1. Grades „es sei ein Leichtes gewesen” abhaengt bzw. ihm untergeordnet ist) und in einem paarige Kommata eingerahmt. Muss die TM in ihrem eigenen Woerterbuch das Wort „die” suchen, wird sie viele Morpheme finden, die alle „die” heissen. Ohne hier auf die vielen weiteren Formen von „die” im Wortfeld einzugehen (inkl. Assoziationen, Wertungen u. v. a.), ergibt auch schon eine hier nur verkuerzte Anwendung der Grammatik, sogar noch weiter verkuerzt, indem hier nur die Artkel aller drei Genera betrachtet werden, 8 unterschiedliche „die”, die zur Auswahl stehen: Zweimal „die” im Maskulinum Plural, Nominativ und Akkusativ. Zweimal „die” im Neutrum Plural, Nominativ und Akkusativ. Viermal „die” im Femininum Singular, Nominativ und Akkusativ und Plural Nominativ und Akkusativ. Nach „ein Leichtes gewesen” koennte der Satz auch weitergehen mit einem Einschub, der z. B. lautet „die Koenigin ist der gleichen Meinung” (wobei „die” ein Nominativ Singular Femininum waere). Oder er koennte mit irgendeinem anderen Satz (also auch mit einem gereihten Nebensatz 1. Grades oder gereihtem Hauptsatz) fortfahren. Die TM kennt bei dem Begriff „die” zwar schon alle theoretischen Moeglichkeiten, weiss aber erst nach Rueckmeldung der folgenden Woerter („”Neue Welt” zu entdecken”), dass es sich bei diesem „die” um einen finiten (bestimmten) Artikel im Akkusativ Femininum Singular handelt.
• (25) „Neue Adjektiv (Teil eines Namens), Femininum, Akkusativ, Singular. Das Adjektiv „Neue deutet bereits auf ein Femininum Singular hin, denn Maskulinum und Neutrum kann dieses Adjektiv auch im Plural wegen des fehlenden n am Wortende (die neuen Maenner, die neuen Haeuser) nicht sein. Dass *Neue” gross geschrieben ist, koennte bedeuten, dass nach Neue kein Nomen mehr folgt, da ein Adjektiv auch zum Substantiv werden kann. Was dagegen spricht, ist das Anfuehrungszeichen, das hinter „Neue nicht beendet wird. Ehe die TM entscheiden muss, ob es sich hier vielleicht um einen Schreibfehler handelt, liefert das folgende Nomen Welt” mit der Beendigung der Anfuehrungszeichen die Information, dass hier kein Schreibfehler vorliegt, sondern dass die beiden Woerter „Neue Welt” zusammengehoeren und dass es sich hier um eine besondere Bezeichnung, vielleicht sogar um einen Eigennamen, handelt. Im uebrigen koennen derart zusammen gehoerende Begriffe im Woerterbuch auch als Wortpaar oder Wortgruppe (vgl. bei Redensarten u. a.) neben der eins zu eins Abspeicherung von Einzelwoertern auch zusaetzlich noch zusammen abgespeichert werden.
• (26) Welt” Nomen (Teil eines Namens), Femininum, Akkusativ, Singular. Sowohl nach dem Wort „Neue, als auch nach dem Wort Welt”, aber auch nach beiden zusammen, ist immer noch der Kasus der „Neue Welt” offen. Theoretisch koennte „die „Neue Welt” auch jetzt noch nach dem Nebensatz („es sei ein Leichtes gewesen”) im Nominativ stehen, wenn naemlich der vorliegende Satzteil nur ein Satzeinschub waere (z. B. „die „Neue Welt” war naemlich gar nicht so neu”). Im Wortfeld spielen sich auch bei dem Morphem „Welt” wieder die gleichen Dinge ab, die schon bei einigen anderen Begriffen dargestellt wurden. Wieder gibt es auch bei diesem Begriff Unterschiede zwischen den gleichen Morphemen „Welt”, die fuer die Sinnerkennung und fuer Uebersetzungen auseinander gehalten werden muessen. Auch hier geben Redensarten, Phraseologien usw. gute Beispiele ab fuer die unterschiedlichen Inhalte des Begriffs „Welt”: a) Wortnaehe: Erde, Kugel, Weltkrieg, Weltsprache, Weltfrieden, Weltbevoelkerung, b) Phraseologien: „Mann von Welt”, „die Welt erschuettern”, der Welt „lebewohl” sagen, Miss Welt, die grosse Welt, die ganze Welt usw. Pro Morphem kann das feminine Nomen ”Welt” 4-mal vorkommen (d. h. im Singular in allen 4 Kasus plus Vokativ).
• (27) zu Bestandteil des Infinitivs („vgl. entdecken”). Neben diesem vorgenannten „zu”, gibt es noch einige Woerter „zu” mit anderen Funktionen (als Adverb, Partikel, Verbpartikel, Praeposition usw.). Auch ohne grammatische Benennung dieser Funktionen leuchtet es ein, dass es sich bei den folgenden Beispielen um verschiedene „zu” handelt: „ab und zu”, „zu leicht”, „der Weg fuehrt zu einer Stadt”, „man schliesst die Tuer zu”, „er hat nichts zu tun” usw.). Es dauert also bis zu diesem vorletzten Wort „zu”, das einen folgenden Infinitiv und die „Neue Welt” im Akkusativ vermuten laesst.
• (28) entdecken, Verb (vgl. „entdecken”). Und es dauert bis zu diesem letzten Wort „entdecken”, einem transitiven Hauptverb, bis an die TM die erste, keine Alternativen mehr duldende, Rueckmeldung kommt, dass es sich bei („Neue Welt”) tatsaechlich um den vermuteten Akkusativ, Singular, Femininum handelt. Es gibt in der (deutschen) Sprache kein einziges Wort, das grammatisch nicht in unverwechselbarer und einmaliger Weise definiert ist. Darum wird bei der Analyse des Quelltextes Sprachgrammatik zugunde gelegt, und nicht die reduzierte und unvollstaendige formale Grammatik. Entsprechend koennen saemtliche moegliche Woerter und Wortformen mit einem eigenen, unverwechselbaren Zahlencode versehen werden, der eine eindeutige grammatische Beschreibung repraesentiert (und spaeter bei Uebersetzungen eins-zu-eins zuordbar ist). Obwohl das Aufbauen eines solchen Lexikons eine bis zu 500-fache Erweiterung des zugrundeliegenden Grundwoerterstamms bedeutet, muessen die zusaetzlich enstehenden neuen Woerter und die grammatischen Unterformen groesstenteils nicht manuell in das hier zu entwickelnde Lexikon eingegeben werden, sondern koennen ueber Algorithmen gebildet werden. Auch Ausnahmen, wie z. B. die irregulaeren Verben, waeren per Algorithmen berechenbar und in das eigene Lexikon eintragbar. Sie stellen allerdings derart kleine Quantitaeten dar, dass Sie manuell in das hier entwickelte Lexikon eingegeben werden. Bei dem eigenen Woerterlexikon handelt es sich um eine relationale Datenbank, die leicht zu verwalten ist, da als Datentyp nur Integers (also ca. 2,1 Milliarden ganze Zahlen) auftauchen. Nur nach dem ersten Schluesselattribut, also der einen Nummer, die im Lexikon einem einzigen Wort zugewiesen ist, erscheint dieses Wort als Character und in Buchstabenschrift, damit die in Normalschrift geschriebenen Woerter des Quelltextes ihre Pendants erkennen koennen. Der Rest, d. h. alle Inhalte der Spalten bestehen nur aus Zahlen, naemlich aus willkuerlich festgelegten Zusatzcodes (die soetwas angeben wie „transitives Verb”, „abstraktes Hauptwort” u. ae.) und aus den Nummerierungen anderer Woerter, die als foreign keys mit dem vorrliegenden Datensatz verknuepft sind, weil sie typischerweise in Quelltexten auftauchen und dem vorliegenden Wort bei seiner grammatischen und anderen Analyse dienen. Bei der eins-zu-eins Aufsplittung eines Vokabulars (d. h. pro Wort gibt es nur einen Sinn) ergeben sich immer nur sehr wenige Doppelbelegungen, von denen mittels dieser im Datensatz aufgefuehrten Zusatz- und Umgebungs-Woerter leicht das richtige, mit dem Quelltext identische Wort, identifiziert werden kann. Selbst die vielen hunderttausend Verben mit ihren ueber 500 Konjugationsformen lassen sich auf ganz normalem Natursprachenniveau mittels „ganz normaler” Natursprachengrammatik (die fuer die mathematischen Sprachlogiker offensichtlich nicht existiert, denn sich ihrer zu bedienen, wurde bis heute noch nicht einmal versucht) „formalisieren”, d. h. numerischen Algorithmen zufuehren. So verursachen Verben, inkl. Passiv, trotz ihrer 500 Konjugationsformen bei der Beugung nur 16 verschiedene Endungen und sind somit in ihrer berechenbaren Ordnung leicht zu verwalten und zu identifizieren.

Because with numbering of all words a complete "formalization" is possible also at this level shown below, which allows numerical algorithms in every area of a text:

• (1) Christoph Nomen (proper name), proper name, probably in the genus Masculinum. If not saved as such, the original word from the source text will be used in the translation. As a proper name of a (mostly human) being, this word is almost always in the number singular and represents an article-less noun, usually without articles, if this noun (mostly in the plural) is not exactly an appelative (ie a generic name). The case is (still) unknown here. Later, this noun becomes a dative.
• (2) Columbus noun (proper name), (proper noun), dative. If not saved as such, the original will be taken over from the source code into the translation. Two proper names in a row make it likely (in German) that they are first names and surnames. Also a singular. Everything else is here as under 1). If every word of the present text (as intended) is connected to every other word of this text about its own word field, and even more so, if already in the storage of this name previously close terms are entered in the word field (including "egg", " America "etc.), the recognition of the identity is additionally secured. The case of the two names is, as I said, still unknown at this point: Since there is usually no article used in proper names, five cases are possible here, vocative ("Christopher Columbus, come here") plus 4 other cases singular: The, the, the, the. (the CK is doing something, the CK house is here, we give something to the CK, we ask the CK). Later in the text, this noun turns out to be a dative.
• (3) is verb (auxiliary verb), 3rd person, singular, Presens (see "held"). After the proper name of Christopher Columbus is still the case of this noun unknown, but it is expected after this noun a verb that appears here in the form of a "will". However: a) This "will" may mean that CK (nominative) z. B. now becomes Vice-king "will". b) This word "will" can (also in the nominative) but also represent a help verb to initiate Futur I and II: CK "will" discover America (or will have discovered it). c) This "will" may also be a passive, with CK also nominative (CK will be rewarded or rewarded). d) It can also be a passive help verb (in 3rd person Sing.), which forms in connection with a still pending participle II of a Vollverbs a so-called. Partizipialphrase and this first main clause (as 1. Wuerfel) concludes, at the CK in the presence or in the future a dative represents (eg the CK is given something or something will be given him).
• (4) after preposition (temporal, with dative) After the above "will" follows as fourth word the word "after" (it is the first word in the second dice). According to the procedure provided here, the above mentioned 3 words ("Christopher Columbus will ...") will remain in waiting position, and the next three smaller parts of the text or "dice" (bits like "after his return from America" and "During a meal with Cardinal Mendoza" and "in the year 1493") are word for word translated one after the other and completed in cubes. These aforementioned dice 2, 3 and 4 are not actants (ie semantically significant sub-nets or phrases), but only in temporal and grammatical juxtaposed increments within the former, their superior, "matrix sentence", and you can quickly find that none of them Words can offer anything to the first main clause ("Christopher Columbus will") in order to complement it and thereby conclude grammatically and semantically. Prepositions such as "to" or "out", among others, do not form part of the sentence but require an addition. But there are many "after" words that are homonymous in German (which does not mean that they have to be in the foreign language): See "sailing to America" for "sailing to America" or "after lunch" for "after" the food "etc.). The supplementation required by the preseposition thus decides first about the type of word "for" (the static local (after the traffic light), purposeful local (sailing to America), temporal (after eating), modal (according to Swiss Art ) can be uva). Only the following noun (often only after his article), to which the Praeposition refers, gives the TM the possibility to select the correct "after" from his Morphem vocabulary. Here, in the present case, the postition follows "after" the possessive article "his" and then the sustantive "return".
• (5) his possessive articles, feminine, dative, singular ("see" return "). As far as the possessive article "his" is concerned, the grammatical masculine of Cristoph Columbus hovers over all parts of the text (that is why "their" is not mentioned here), but the cases are of course determined by the dative singular of "return". therefore, the present text refers to 'after his return'.
• (6) Return Noun, Feminine, Dative, Singular This interpretation is simple because the noun "return" is unique (ie, there are no ambiguities or double meanings of that word). "Return" is here temporal, ie it forwards the temporal "after" (and not the local one, for example). Also this temporal "after" demands a dating from its addition. This dative is not to be regarded as the noun "return", since "return" exists only in the singular and it has no inflectional morpheme in the dative singular. All cases of this word (which is a conversion of the verb "to return") are always called "return", without a single suffix at the end. For each morpheme, the feminine noun "return" can occur 4 times (ie in all 4 cases in the singular).
• (7) from Praeposition (local with Dative) "from America" is also a Praeposition ("out") and its addition ("America"), a local proper name, according to mostly without articles and always in the singular. As for the initial three words ("after his return"), they should be completed, because the word "from" clearly introduces a new statement, so that it does not have to be expected that the three seemingly coherent words "follow his Return "must change again in their mind, because subsequent terms require a different interpretation of these three morphemes. Also "off" has (like "after") many meanings ("accidentally", "out of fashion", "out of memory", "done out", "I ask it out", "out of wood", " from America "etc.). So it must be clarified beforehand for the choice of the right word "out" by the TM, what kind of word "out" it is.
• (8) America noun (proper name), neuter, dative, singular. Again, only the following noun provides information. (Proper noun). "America" needs a local "out" that requires a dative, ie America is in the dative. This dative is difficult to recognize because America has no suffix in the dative singular (just like the "return") and (unlike the "return") has no article. But the latter is irrelevant, since the two words given here ("from America") represent a very simple arrangement of words and the equally simple arrangement of the first three words ("after his return") does not "get in the way". This is the second dice "after his return from America" also completed.
• (9) during praeposition (local with genitive) The word "during" is not a primal position. If the sentence in the text "while they eat," would be "during" only a so-called subjunction, which connects a dependent subset (subordinate clause) with the main clause. As with the prefixes "after" and "out", a noun whose case (as has already been stated) is always determined by the respective primal position belongs to the primal position. Here it is the genitive (less accurate but sometimes the dative is used here).
• (10) an article (indefinite), neuter, genitive, singular ("compare food"). Indefinite article, grammatically similar form as "food" (genitive, singular, neuter).
• (11) Essen's noun, neuter, genitive, singular, with suffix. Between the noun "Essens" (genitive, singular, neuter) and the preposition "while" (formed from the verb "essen") stands the infinite article of the noun "one" (also in the genitive, singular, neuter). For each morpheme, the neutral noun "food" can occur 3 times (ie in 3 cases in the singular).
• (12) at Praeposition, (local with Dative) Also "at" is again a local Praeposition, which calls for the associated noun the Normal Casal Dative and here indicates a space or a location.
• (13) cardinal noun, masculine, dative, singular (to "Mandoza") "Cardinal" is a title that usually requires an article ("Cardinal" or "Cardinal"), but here can be part of the proper name "Mendoza" without an article.
• (14) Mandoza noun, masculine, dative, singular (to "cardinal") Since neither the proper name nor the title show their dating by article or word ending, this dative is invisible and can not be returned to the preposition "at" , (Proper name), masculine, dative, singular. Nevertheless, the TM will accept this uncertainty, as there is nothing to suggest that Cardinal Mendoza is a dative (singular, masculine).
• (15) in the pre-position (here temporal with dative) Also, the following 3 words "in 1493" are just another insert and take grammatically no relation to the first 3 words. These 3 following words begin with the prefix "im" (where an "in" is merged with the following definite article "the", which can happen in abstracts and above all phrases (the so-called fixed phrases)), ie a preadposition, for static statements of the neutral "year" the dative ("in the (in) the year") and for directed statements the accusative (until the (in) the year ") demand.
• (16) years Noun, neuter, dative, singular, with suffix Instead of "in the year 1493" not only the fusion to "in" has prevailed, but also the formulation "in the year". In fact, the suffix "e" after "year" is as unnecessary as the whole phrase "in the year" or "in the year", it would also suffice to indicate only the year. But, as I said, this is a so-called "fixed turn", which can be stored as such in addition to the individual morphemes "in" and "years" in the dictionary. After "in the year" follows a vacancy for an explanation or a year.
• (17) 1493 Number (year) The number "1493" does not have to be defined either for the recognition of the meaning or for translations.
• (18), verb, (participle II), (see "will"). After 13 morphemes and a number, all of which were individually and together not suitable to complete the initial main clause ("Christopher Columbus will ..."), but appears at position 18, the word "held, with this main verb" held The initial main clause finally completed and now decides on the case of CK at the very beginning. Christopher Columbus is there, as we now know after the 15th Word, in the dative, because he is "held up". Provision has two meanings, ie the CK is held a weapon, ie a concrete object, or the CK (as is the case here) something abstract held up. But this distinction is grammatically irrelevant. If the CK has a weapon or betrayal, in both cases the sentence can be ended with one of these two accusatives. But the absence of an accusative creates a so-called vacancy, that is, it makes an explanation of what the CK is held up. This results from the fact that the actant "(the) CK is held up"), although his own dice may be, but is still grammatically unfinished. Both could, as mentioned above, happen briefly with an accusative: The CK is held a weapon or a betrayal (point, main sentence completed). Or this expected accusative stays off, then an explanation is expected instead (which explains what the CK is held up). This is initiated with "that" or takes place in direct or indirect speech. The explanation follows in this text with the words "it was easy" (ie in the subjunctive of indirect speech), which then creates another vacancy by waiting for a minor subordinate clause, which in turn explains what exactly "easy This interpretation is simple, because the noun "return" is unique (ie there are no ambiguities or double meanings of that word). "Return" is here temporal, ie it fodert the temporal "after". Also the temporal "after" demands a dating from its addition. This dative is not to be regarded as the noun "return", since "return" exists only in the singular and it has no inflectional morpheme in the dative singular. All cases of this word (which is a conversion of the verb "to return") are always called "return", without a single suffix at the end.
• (19) it pronouns Following the verb of the main clause (Praesens Indicative Passive) and before this "it" (with which a new clause begins, there is a comma that (since in the main clause no noun object for the pending statement was present) with "that" , with direct speech or (as here) with indirect speech, the explanation provides what the CK is held, this is a different "it" than, for example, in "it is me bad," a sentence in The "it" could easily be omitted ("I am bad"), because this "it" is a so-called placeholder and fulfills a function as a subject. "It was easy" (or simply: It was easy) the "it" must not be missing, because exactly this accusative (it, this, etc., which is explained in the following even more, namely with "the new world" to discover ") as a subject belongs to the sentence (a subordinate clause of the first degree, grammatically in the subjunctive II de indirect speech directly follows the main clause (CK is held ").
• (20) be Verb (auxiliary verb), 3 person, subjunctive The subordinate clause 1st degree "it was easy" here used the indirect speech with Referativesjunjunktiv II. Without grammatical errors, the auxiliary time word "been" could also be omitted, the statement would have then, also grammatically correct, stated, "it would be easy to discover the" New World ". However, it is more semantically accurate to add the word "been", since there is no possibility of repetition in the present and future. Because the "New World" could only be discovered once.
• (21) an article (indefinite), neuter, nominative, singular, (see "easy"). This is not a number, a cardinal number adjective, etc., but an indefinite, infinite article. Even an indefinite article is mitdekliniert mit the associated noun. Since the noun is not yet known at this point, the word "a" could occur either as a masculine nominative singular (a man) or as a neuter nominative singular (a house) or as a neuter singular accusative (a house).
• (22) Light noun, neuter, nominative, singular (see "on"). This is a substantivized adjective in the nominative singular. For the TM to be developed here, it is logical to translate this "an easy one", although with almost the same meaning, only the adjective "simple" would have sufficed.
• (23), verb (participle II), infinitive, with word stammering and praefix. As stated under "be," this subjunctive is a subjunctive II, that is, a preterite of "being," in which this "being" is used as a finite verb ("being" as a noun) and its perfect, pluperfect and Futur II with "been"(it's been, it had been, it will have been).
• (24) the articles (finite), feminine, accusative, singular. The previous word "been" and this next word "the" are correctly separated from a comma. The definitions developed in the TM, which are only described here in this Annex. T. were presented with some grammatical rules, are so compelling that they work, even with wrong or completely missing punctuation! The only drawback would be in case of false or missing punctuation only that the sense control would have finished their work only at the very end of the sentence, if all words of the sentence were recognized as correct (or not) despite the lack of punctuation. The whole set of sentences now to discover the "New World" is a subordinate clause of the second degree (which is dependent on the previous subordinate clause "it was easy" or inferior to it) and framed in a pair of commas , If the TM has to search for the word "the" in its own dictionary, it will find many morphemes, which are all called "the". Without going into the many other forms of "the" in the word field (including associations, valuations, etc.), even here a shortened application of the grammar results, even further shortened, by considering only the articles of all three genera, 8 different "the" to choose from: Twice "the" in the masculine plural, nominative and accusative. Twice "the" in the neuter plural, nominative and accusative. Four times "the" in the feminine singular, nominative and accusative and plural nominative and accusative. After "been easy", the sentence could go on with a slot that z. For example, "the queen is of the same opinion" (where "the" would be a nominative singular feminine). Or he could proceed with any other sentence (ie also with a ranked subordinate clause of 1st degree or ranked main clause). The TM knows the term "the" already all theoretical possibilities, but knows only after feedback of the following words ("" new world "to discover"), that this "the" to a finite (certain) article in the Accusative Femininum Singular acts.
• (25) "New adjective (part of a name), feminine, accusative, singular. The adjective "new" suggests already a feminine singular, because masculine and neuter, this adjective can not be in the plural because of the missing n at the end of the word (the new men, the new houses). The fact that "new" is capitalized could mean that no nouns follow new ones, since an adjective can also become a noun. What speaks against it, is the quotation mark, which is not terminated behind "new ones. Before the TM has to decide whether this is possibly a spelling error, the following noun world "with the completion of the quotes provides the information that there is no spelling error, but that the two words" New World "belong together and that it is this is a special name, maybe even a proper name. Incidentally, such terms belonging together in the dictionary can also as a word pair or phrase (see in phrases, etc.) in addition to the one-to-one storage of individual words also additionally stored together.
• (26) World "noun (part of a name), feminine, accusative, singular. Both after the word "new and after the word world", but also after both together, the case of the "New World" is still open. Theoretically, "the" New World "could still stand in the nominative after the subordinate clause (" it was easy "), if the present sentence was only a sentence insert (eg" the "New World" was indeed not so new "). In the word field, the morphemes "world" play the same things again, which have already been shown in some other terms. Again, there are also differences in this term between the same morphemes "world", which must be kept apart for the recognition of meaning and for translations. Here, too, phrases, phraseologies, etc., give good examples of the different contents of the word "world": a) word meaning: earth, globe, world war, world language, world peace, world population, b) phraseologies: "man of the world", "the world shattering ", say goodbye to the world, Miss world, the big world, the whole world, etc. Per morphem, the feminine noun" world "can occur 4 times (ie in the singular in all 4 case plus vocative).
• (27) as part of the infinitive ("see discover"). In addition to the aforementioned "to", there are still some words "to" with other functions (as adverb, particle, verb particle, preposition, etc.). Even without a grammatical naming of these functions, it is clear that the following examples are different "to": "now and then", "too easy", "the way leads to a city", "you close the door "," He has nothing to do "etc.). So it takes until this penultimate word "to" that suggests a following infinitive and the "New World" in the accusative.
• (28) discover verb (see "discover"). And it takes until this last word "discover", a transitive main verb, until the TM comes the first, no longer tolerating alternatives, that ("New World") in fact the suspected accusative, singular, feminine is. There is not a single word in (German) language that is grammatically not defined in an unmistakable and unique way. This is why the analysis of the source text employs speech grammar, not the reduced and incomplete formal grammar. Correspondingly, all possible words and word forms can be provided with their own, unique numerical code, which represents a clear grammatical description (and can later be assigned to one-to-one translations). Although the construction of such a lexicon is up to a 500-fold extension of the underlying root word stem, the additional emerging new words and the grammatical subforms are for the most part not manually entered into the lexicon to be developed here, but can be formed using algorithms. Also exceptions, such. As the irregular verbs, would be calculable by algorithms and entered into your own lexicon. However, they represent such small quantities that they are manually entered into the lexicon developed here. Its own dictionary of words is a relational database that is easy to manage because only integers (ie, about 2.1 billion integers) appear as a data type. Only after the first key attribute, ie the one number, which is assigned to a single word in the lexicon, does this word appear as a character and in alphabetical order, so that the written words of the source text can recognize their counterparts. The remainder, ie all contents of the columns consist only of numbers, namely of arbitrary additional codes (which specify something like "transitive verb", "abstract noun" and others) and from the numbering of other words that are used as foreign keys with the existing data set, because they typically appear in source texts and serve the present word in its grammatical and other analysis. In the one-to-one splitting of a vocabulary (ie there is only one meaning per word), there are always very few double assignments, of which the additional and environment words listed in the data record are easily the correct word that is identical to the source text , can be identified. Even the many hundred thousand verbs with their more than 500 forms of conjugation can be "formalized" at a normal natural language level by means of "quite normal" natural language grammar (which obviously does not exist for the mathematical linguists, because it has not even been attempted yet) ie feed numerical algorithms. Thus verbs, including passive ones, despite their 500 forms of conjugation in diffraction, only cause 16 different endings and are therefore easy to manage and identify in their predictable order.

Annex 2

Regular verbs as an example, as analyzed with a natural language grammar and prepared for algorithmic processing:
Here are the normal, regular, regular, weak verbs, eg. For example, the verb "love" that is transitive. (Intransitive weak verbs are conjugated in the same way, eg "live" do not form passive forms):
Most verbs are weak, ie regular, that is, they do not change their stem during inflection. Most of these verbs end in "infinite". The algorithm captures all word stagnation, optically the infinitives are without a suffix (usually without "en").

• –) Verben sind darin fast immer als solche gekennzeichnet (vgl. Hinweise wie ”V” oder ”v”).
• –) Hier in diesem Beispiel interessieren diejenigen, die auf ”en” enden und diejenigen,
• –) die in den vorgenannten Quellen als regelmaessig (meist mit ”reg.” beschrieben) bezeichnet werden. (Sollten nur die unregelmaessigen Verben (und zwar mit ”irr.”) gekennzeichnet sein, sind diejenigen Verben regulaer, die nicht als irregulaer gekennzeichnet sind.

Since both man and machine do not look at a word stem before, whether it will change in flexing, there is no rule according to which a selection algorithm could be made in advance. The differentiation of these verbs is best done by taking the appropriate notes from a dictionary or foreign-language lexicon or both.

• -) Verbs are almost always marked as such (see notes such as "V" or "v").
• -) Here in this example, those who end in "en" and those who are interested
• -) which are referred to in the aforementioned sources as being regular (usually described as "reg."). (If only the irregular verbs (namely with "err.") Are marked, those verbs are regular, which are not marked as irregular.

With all these verbs mentioned above, the algorithms always proceed in the same way: If one wants to use the system used here (ie the conjugate tables compiled here) (the type and order, however, how the conjunctions are put together is at this point It does not matter that all grammatical possibilities are recorded, and always exactly once per case), this system looks like this for all verbs mentioned above.

For these many hundreds of thousands of verbs, the grammatical forms listed below emerge uniformly, the regularity and predictability of which make it possible to build up the universally available basic lexemes into a morpheme lexicon by means of a few dozen of their own algorithms.

In the case of the weak verbs (live, love, etc.) there are around 540 forms of conjugation (about 745 minus 200 declinations minus clipping) around the root word (infinitive without "en") .Together, these only cause these 540 forms of conjugation 16 different endings (only in the past participle II, besides the ending "t", there is also a prefix, which, as with the regular verbs, is "ge"):
1 time "e!" (Imperative)
1 time "en!" (Imperative)
1 time "t!" (Imperative)
1 time "you" (imperative)
1 time "end" (past participle I)
1 time "st"
1 time "est"
1 time "et"
5 times "t"
6 times "e"
7 times "ten"
7 times "en"
10 times "te"
10 times "test"
31 times "tet"
551 times "ge" (word stem) "te"

In connection with the surrounding words of the grammatical references (see "we" are "loved"), they are always clearly identifiable at the end of algorithms in all cases, including multiple occurrences.

Claims (2)

Protection is claimed for the method of further increasing the natural languages for machine computability, to the extent that each word represents only one sense (as do integers.) While a 2.9 may approximate a 3, it does It is not an integer "three" is defined as a "3" without a mantissa, so a 3.1 is not an integer "3", just as a "2.9" is not a "3". The aforementioned term "every word" means not only the basic words, but every single word, including the subordinate and minor forms of a root word, as it is, for. B. occurs in grammatical diffractions. Accordingly, "house" (the second case of house) is a separate word, but also "house" in the third case and "house" in the fourth case. If this aforementioned idea had been addressed before, then instead of a dictionary with one million basic words, approximately 500 such books would have had to be printed for all secondary forms. This could hardly be handled, which is why it has never happened before. Today, such a derailment of vocabularies but can be detected with computers and represented by algorithms. For this protection claim is logically the protection for the self-developed word fields and especially numbering. Even if a numbering does not protect itself, it should be protected with the words (and their meaning), because here words and numbers belong together. From this principle, described under 1), it is possible to have native-language texts (that is, translated into no foreign language) automatically (that is, algorithms) examined for their grammatical and in part meaningful consistency To translate these texts into numbers using a (back) translation into their own (ie non-foreign) language in order to communicate their grammatical and other functions to any compatible receiver that has the same number assignment. This recipient may be a foreign-language lexicon that is "out of the office" or it may be part of the same program that contains not just one vocabulary but two analog vocabularies of two different languages. If a finished translation of "HANA I" is delivered, it can be linguistically modified with the technique of the other translating machines. In this respect, "HANA I" can also improve the quality in terms of sense and accuracy for other translation machines with "subcontracting services".